Comparing methods of identifying plants visited by pollinators through morphological and genetic techniques : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Zoology, School of Food Technology and Natural Sciences, Massey University, Palmerston North, New Zealand

Abstract

Interest in bumblebee pollination is increasing as honeybee populations appear insufficient to meet agricultural pollination demand. Effective use of non-honeybee pollinators like bumblebees will also save farmers money, compared to managing commercial honeybees. Potentially, bumblebee populations can be increased by providing supplementary sources of nutrients (plants that blossom outside of crop flowering) on farms. To identify the most beneficial plant species, three methods were proposed that had been used to answer similar questions previously: (1) Using morphology of intentionally and unintentionally collected external pollen from the bumblebees to identify the plants they visited. (2) Using amplicon DNA sequencing (metabarcoding) on DNA extracted from pollen from bees to identify the plants they visited by comparing sequences against reference databases. (3) Using the same amplicon DNA sequencing method on the DNA extracted from bee stomach contents to identify the plants they visited. These methods were compared using five bumblebees (Bombus terrestris), two native bees (Leioproctus monticola) and one honeybee (Apis mellifera). For method-1. 3627 individual pollen granules were counted and identified to 10 plant genera. For method-2 & 3 the target amplification region used for was the choloroplast gene trnL. 2,631,048 reads were obtained from seven pollen samples and for method-3. 916,868 reads were obtained from eight stomach DNA samples. 732 unique cpDNA sequences were recorded, however many of these referenced to the same genera, or to non-viridiplantae genera or did not reliably match at all (low match and coverage). After the bioinformatics, 35 New Zealand plant genera were identified although 13 represented pollen and stomach exclusive genera, which contained only one genus with >1% which was 1.7%. These findings resulted in the claim that the plant genera visited for pollen does match the plant genera visited for nectar. The main outcome of this study is that pollen DNA amplicon method was the most effective in terms of time and quality, and can be trusted to be representative of the species bumblebees feed on, as the third main finding showed using pollen did not exclude any significant genera only found in the stomach, or include significant genera that bumblebees do not consume. However, the failure of the metabarcoding approach to detect wild radish pollen suggests this approach is not perfect.